PURPOSE: The aim of this study was to describe if and when a sample of traumatic brain injury (TBI) patients could finish their time of sick leave during a 4-year follow up and to explore which factors that influenced the time for sick leave. MATERIALS AND METHODS: All persons, 1999-2002, between 18 and 64 years of age (250 in total), admitted to the emergency room and diagnosed according to ICD 10 as S062 and S063, were included. Demographic data were gathered from medical charts and data concerning sick leave 1 year before trauma and 4 years after trauma, were collected from the Swedish social insurance system. To explore predictors of sick leave, two logistic regressions were performed. RESULTS: The sample (mean age 39.68) consisted of 78% men. More than half of the accidents were due to fall. In the sample, 28 % was on sick leave on the day of trauma and 96 % of these were still on sick leave 4 years after trauma, compared with 39 % in the group not on sick leave on the day of TBI. Sick leave at the day for trauma was found to be a predictor for sick leave 4 years after trauma for the whole group (p = 0.000) together with Glasgow Coma Scale (GCS) (p = 0.002) and length of stay (p = 0.049). In the logistic regression with only the group not on sick leave, the only significant variable was GCS (p = 0.003). CONCLUSION: The findings support the necessity to consider premorbid and social factors in the TBI rehabilitation.